Note: This post was delayed a couple weeks.

Today Scott Stults and myself traveled up to DC to exhibit and attend the DoD Open Technologies Conference. We attended last year where Matt Jenks and myself spoke on Agile techniques for Green Field/Brown Field Enviroments. The conference has moved to the Ronald Reagan International Trade Center in downtown DC, and as a venue is a much nicer, much brighter space.

The initial keynote was by Bill Vass, President and COO, Sun Microsystems Federal, and I expected something similar to Scott McNeely’s keynote at FOSE. His presentation touched on the same topics of why Sun has open sourced it’s IP, but focused a bit more on the US Government’s involvement. One of the key strengths that Bill talked about was the ability for the Government to influence open source projects, and he gave a couple examples. One of the ones that he gave was that the NSA wanted to add “TE”, Type Enforcement, to Solaris. They asked Sun to do it, and Sun wasn’t interested because Sun’s own internal security engineers didn’t feel that TE was part of the vision for security. So what did the NSA do? They funded an open source project that extended Solaris to support TE, possible only because of Solaris being open source! And lo and behold, time passed and Sun ended up taking that enhancement and rolling it into their supported version of Solaris. There were a number of other examples ranging from changes to Solaris to changes to ODF made by various government agencies possible only because of open source. Without the source being open, even if the Government issued an RFP to a commercial vender to make these changes, without access to the source it wouldn’t be possible.

Bill also put out some statistics on Sun as the #1 contributor to open source, which, while true, is mostly because of them taking so much proprietary code and open sourcing it, versus starting projects from open source. Sun’s revenue from licenses has gone up drastically since open sourcing their stack, but I feel that is because people were not willing to pay, post the dot com era, for Solaris when Linux was available. But with Solaris being on the same pricing structure as Linux, then it is more compelling!

He also made the same plug that Scott did that Java is on 6 billion devices. I hate that statistic, because when someone hears that, they mentally think of a rich environment where devlopers can write an app and have it run everywhere! But what that really means is that 5.5 billion of those devices are cheap cell phones running various versions of J2ME. I’ve done some hacking with J2ME, and no two phones support it the same way, and there are too many levels of J2ME compliance. J2ME really is “Write Once, Test Everywhere!”

He pointed out that open source has many very compelling solutions in the IT plumbing realm of storage, server virtualization, os, database. However, I think open source still has a way to go up the food chain from making a big data center more efficient. He acknowledged that issue, talking about how moving to open source solutions for J2EE and database SQL centric apps is much easier then more traditional, older style apps like an ERP. J2EE, for all the shortcomings of JSP, EJB, etc, does make it easier to move from a commercial app server to something like JBoss or Glassfish by providing very open standards.

He also talked a bit about moving back office functions like calendering and email off of Exchange. “It’s hidden from the user, they still use Outlook”. While I like the message, I don’t see it yet. Exchange and Outlook work very closely together, and people’s email and calendering are very hot button issues.

Lastly he summed up Sun’s Lessons Learned after moving to open source in such a big way:

• More Secure
• Cost less, but not free
• Watch out for putting up your own dev shop, leave customizations to vendors
• Indemnity provides the same protection as a license. This sounded like another plug for Sun’s version of the projects!!!
• Stay with open standards, they make open source work. TCP/IP is critical to use open source networking tools for example. Standardizing on AppleTalk precludes the option of open source solutions!

It was a good presentation, I liked the focus on government helped the message resonate with the audience. However I felt like the focus on desktop’s moving to open source was a bit of a reach. There is so many places on the server, network end of things that open source can help, that changing the desktop is still a long way away. Unless you are building out a pure kiosk type solution, then the advantages that Microsoft enjoys on the desktop are still too strong. However, moving more to the web will reduce that advantage over time!

FOSE 2008 is over, and we’re all back home. Here are my observations from the final day of FOSE 2008:

• It would have been nice to have our people attending more sessions. Arin Sime got to attend a couple of sessions, and Eric Pugh saw the opening keynote speech, but it would have been nice to attend more sessions to get the government point of view of information technology.
• The attendees have a general distrust of custom solutions. It’s not a wonder. They probably suffer from vendor lock, unresponsive clients, and have generally been burned by bad process.
• There are pockets of Agile development, but not many. This probably leads to the problem cited above, but it also reflects a greater problem…
• Contracting officers don’t know how to price Agile development. The most common pricing scheme for development projects is fixed price. Agile can only deliver a fixed price for labor. Maybe it will be easier under a blanket purchase agreement or with sprint-sized task orders.

It’s too early to say whether this was a worthwhile trade show for us. Each day, we definitely had conversations where people had specific problems that they were trying to solve which our team and process could help them solve. Whether or not we can get from problem identification and solution provider acknowledgment to actually providing the solution remains to be seen.

The people who came by our booth yesterday seemed to represent a slightly different mindset than what we saw in Day 1. I can still categorize the attendees:

• People on a mission: They know what they want, and they’re very specific about what they are looking for. Questions like “Can you do J2EE” and “Are you familiar with Drupal [Alfresco, Joomla, Expression Engine]?” were much more common in day 2 than they were in day 1.
• Browsers who either haven’t heard of open source or misunderstand its meaning: These are fine people to talk to, as there is an opportunity for us to tell them what it is and what it isn’t. Firstly, let me say that open source is not the answer to everything (saying something like that at OSCON would get me lynched. As an aside, I’d wager that the number of people who attend both FOSE and OSCON represents less than 1% of either conference’s attendee population. That number disappoints me greatly, as both populations need to learn about each other over the next few years if our government is to remain competitive without spending greater and greater amounts of a percentage of GDP.
• Schwag runners: The group of people almost sprinting through the exhibit hall sort of reminds me of the Filene’s basement sale. It reinforces why we don’t give out schwag. Perhaps the companies who give out schwag wind up having more conversations with people that they would not have otherwise had, but watching the surreptitious glances of candy grabbers hoping that nobody from the booth actually talks to them makes me think otherwise.

The one group of people whom I expected to see some of that I have yet to see is contracting officers. I know that many agencies are not meeting their set-aside goals, so I would have expected the KOs to be swarming the small business pavilion to find the vendors who can help them meet their goals. We have seen a few primes come by and talk to us and when they discovered that we are a service-disabled veteran-owned small business (SDVOSB), then the timbre of the discussion changed (though we’d rather have people be interested in working with us because of the quality of work that we do) . Maybe, as someone put it, the KOs have been burned too badly by low quality set-aside qualifying firms who can’t actually do the job in the past that they are now looking at the major primes to serve as the vetting mechanism. It’s not a bad idea.

I’m also surprised that nobody is blogging about FOSE. The number of blog posts found on Technorati covering FOSE takes up less than a page. Comparatively, OSCON is usually a top (if not the number one) search term on Technorati during OSCON. There are about 45 news articles on Google covering FOSE, but I would still expect people to be writing (and blogging) more about this convention. It’s huge. There must be at least 4 times as many people here than at OSCON. At OSCON, crowds come in waves as people come out of sessions. At FOSE, there is a constant stream.

Still, being at FOSE is worth it, because it’s important that government users of software become more acquainted and comfortable with open source products and don’t have the not invented here (or by a small group of proprietary software providers) syndrome. There are pockets of users of open source products, and some who are interested in learning more, but the road ahead is a long one. We intend to be there; as citizens, we have a vested interest in how our taxpayer dollars are spent.